Creating a project
A project is the access shell around one or more governed modules. Its modules may contain tabular data, source documents, or both.
Before you start
You need:
- Add Projects permission.
- At least one ready module.
- A clear project name and purpose.
- A decision about who should have access.
If no modules are available, use Add Data first.
Create a project
- Open Projects under Workspace.
- Select Create Project.
- Enter the project details.
- Select one or more available modules.
- Choose All Users or Restricted Access.
- Review the effective LLM exposure.
- Select Save.
File-only Add Data journeys open this blade automatically with their new module selected.
Choose modules
Each module shows its kind, status, and LLM exposure level. Ready file modules may still show Vectorising or Partial while their individual documents finish processing.
When modules have different exposure levels, the most restrictive level governs the project.
Control access
All Users
Every portal user in the organisation may open the project, subject to role permissions.
Restricted Access
Select the individual users who should have access. The project creator remains an owner, and the Customer Super Admin can access all projects in the organisation.
Project access also governs documents associated with its datasets:
- authorised users may question AI-enabled associated files;
- removing project access removes dataset-derived file access at query time;
- dataset access does not automatically permit downloading the original file;
- no synthetic personal share is created.
What opens when you select a project
| Project capability | Default destination |
|---|---|
| File modules only | RebelCore Agent |
| Any tabular or mixed module | Project Tree |
Tree remains the curation and semantic-training environment for tabular data. A file-only project skips it because its documents are already prepared for semantic retrieval.
Edit a project
You can update project details and access membership. The selected modules are shown for reference and remain fixed.
Next steps
- For tabular or mixed projects, open the Tree and build the required semantic dataset.
- For file-only projects, open Agent and select the available file dataset.
- Review LLM exposure levels.